A new study by the European Broadcasting Union (EBU) and BBC has evaluated how accurately AI systems like ChatGPT, Google Gemini, Microsoft Copilot and Perplexity generate news summaries.
The research found that most AI-generated responses contained some errors, with Google Gemini performing notably worse — often failing to cite sources, mixing satire with real news and misquoting information. Although all systems showed improvements over six months, Gemini still lagged behind its peers.
Despite these issues, 42% of UK adults said they trust AI’s accuracy, and that number rises among younger users — even though 84% claim factual errors would damage their trust. The findings raise serious questions about relying on AI for news.
Source: Android Authority
